Mercurial > repos > aaronquinlan > multi_intersect

comparison BEDTools-Version-2.14.3/src/mergeBed/mergeMain.cpp @ 0:dfcd8b6c1bda

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Uploaded
author aaronquinlan
date Thu, 03 Nov 2011 10:25:04 -0400
parents
children
comparison
equal deleted inserted replaced
-1:000000000000 0:dfcd8b6c1bda
1 /*****************************************************************************
2 mergeMain.cpp
3
4 (c) 2009 - Aaron Quinlan
5 Hall Laboratory
6 Department of Biochemistry and Molecular Genetics
7 University of Virginia
8 aaronquinlan@gmail.com
9
10 Licenced under the GNU General Public License 2.0 license.
11 ******************************************************************************/
12 #include "mergeBed.h"
13 #include "version.h"
14
15 using namespace std;
16
17 // define our program name
18 #define PROGRAM_NAME "mergeBed"
19
20
21 // define our parameter checking macro
22 #define PARAMETER_CHECK(param, paramLen, actualLen) (strncmp(argv[i], param, min(actualLen, paramLen))== 0) && (actualLen == paramLen)
23
24 // function declarations
25 void ShowHelp(void);
26
27 int main(int argc, char* argv[]) {
28
29 // our configuration variables
30 bool showHelp = false;
31
32 // input files
33 string bedFile = "stdin";
34 int maxDistance = 0;
35 string scoreOp = "";
36
37 // input arguments
38 bool haveBed = true;
39 bool numEntries = false;
40 bool haveMaxDistance = false;
41 bool forceStrand = false;
42 bool reportNames = false;
43 bool reportScores = false;
44
45 for(int i = 1; i < argc; i++) {
46 int parameterLength = (int)strlen(argv[i]);
47
48 if((PARAMETER_CHECK("-h", 2, parameterLength)) ||
49 (PARAMETER_CHECK("--help", 5, parameterLength))) {
50 showHelp = true;
51 }
52 }
53
54 if(showHelp) ShowHelp();
55
56 // do some parsing (all of these parameters require 2 strings)
57 for(int i = 1; i < argc; i++) {
58
59 int parameterLength = (int)strlen(argv[i]);
60
61 if(PARAMETER_CHECK("-i", 2, parameterLength)) {
62 if ((i+1) < argc) {
63 bedFile = argv[i + 1];
64 i++;
65 }
66 }
67 else if(PARAMETER_CHECK("-n", 2, parameterLength)) {
68 numEntries = true;
69 }
70 else if(PARAMETER_CHECK("-d", 2, parameterLength)) {
71 if ((i+1) < argc) {
72 haveMaxDistance = true;
73 maxDistance = atoi(argv[i + 1]);
74 i++;
75 }
76 }
77 else if (PARAMETER_CHECK("-s", 2, parameterLength)) {
78 forceStrand = true;
79 }
80 else if (PARAMETER_CHECK("-nms", 4, parameterLength)) {
81 reportNames = true;
82 }
83 else if (PARAMETER_CHECK("-scores", 7, parameterLength)) {
84 reportScores = true;
85 if ((i+1) < argc) {
86 scoreOp = argv[i + 1];
87 i++;
88 }
89 }
90 else {
91 cerr << endl << "*****ERROR: Unrecognized parameter: " << argv[i] << " *****" << endl << endl;
92 showHelp = true;
93 }
94 }
95
96 // make sure we have both input files
97 if (!haveBed) {
98 cerr << endl << "*****" << endl << "*****ERROR: Need -i BED file. " << endl << "*****" << endl;
99 showHelp = true;
100 }
101 if (reportNames && numEntries) {
102 cerr << endl << "*****" << endl << "*****ERROR: Request either -n OR -nms, not both." << endl << "*****" << endl;
103 showHelp = true;
104 }
105 if ((reportScores == true) && (scoreOp != "sum") && (scoreOp != "max") && (scoreOp != "min") && (scoreOp != "mean") &&
106 (scoreOp != "mode") && (scoreOp != "median") && (scoreOp != "antimode") && (scoreOp != "collapse"))
107 {
108 cerr << endl << "*****" << endl << "*****ERROR: Invalid scoreOp selection \"" << scoreOp << endl << "\" *****" << endl;
109 showHelp = true;
110 }
111
112 if (!showHelp) {
113 BedMerge *bm = new BedMerge(bedFile, numEntries, maxDistance, forceStrand, reportNames, reportScores, scoreOp);
114 delete bm;
115 return 0;
116 }
117 else {
118 ShowHelp();
119 }
120 }
121
122 void ShowHelp(void) {
123
124 cerr << endl << "Program: " << PROGRAM_NAME << " (v" << VERSION << ")" << endl;
125
126 cerr << "Author: Aaron Quinlan (aaronquinlan@gmail.com)" << endl;
127
128 cerr << "Summary: Merges overlapping BED/GFF/VCF entries into a single interval." << endl << endl;
129
130 cerr << "Usage: " << PROGRAM_NAME << " [OPTIONS] -i <bed/gff/vcf>" << endl << endl;
131
132 cerr << "Options: " << endl;
133 cerr << "\t-s\t" << "Force strandedness. That is, only merge features" << endl;
134 cerr << "\t\tthat are the same strand." << endl;
135 cerr << "\t\t- By default, merging is done without respect to strand." << endl << endl;
136
137 cerr << "\t-n\t" << "Report the number of BED entries that were merged." << endl;
138 cerr << "\t\t- Note: \"1\" is reported if no merging occurred." << endl << endl;
139
140
141 cerr << "\t-d\t" << "Maximum distance between features allowed for features" << endl;
142 cerr << "\t\tto be merged." << endl;
143 cerr << "\t\t- Def. 0. That is, overlapping & book-ended features are merged." << endl;
144 cerr << "\t\t- (INTEGER)" << endl << endl;
145
146 cerr << "\t-nms\t" << "Report the names of the merged features separated by semicolons." << endl << endl;
147
148 cerr << "\t-scores\t" << "Report the scores of the merged features. Specify one of " << endl;
149 cerr << "\t\tthe following options for reporting scores:" << endl;
150 cerr << "\t\t sum, min, max," << endl;
151 cerr << "\t\t mean, median, mode, antimode," << endl;
152 cerr << "\t\t collapse (i.e., print a semicolon-separated list)," << endl;
153 cerr << "\t\t- (INTEGER)" << endl << endl;
154
155 cerr << "Notes: " << endl;
156 cerr << "\t(1) All output, regardless of input type (e.g., GFF or VCF)" << endl;
157 cerr << "\t will in BED format with zero-based starts" << endl << endl;
158
159
160 // end the program here
161 exit(1);
162
163 }